Re: [htdig-dev] indentation?

[Wolfgang M. <Wol...@cu...>]

Hi, Gilles,
Thanks for your fast reply and for saving me from trouble.
> Before you go stripping out code from htsearch and end up having to
> maintain another search program in parallel to htsearch, be sure you
> fully investigate the template capabilities of htsearch.  There's almost

OK. In fact, I stumbled across these things *after* having had started to 
strip htdig.

> nothing in the output that you can change through some template file
> and/or config attribute.  In fact, I recently committed to the 3.1.6
> development code a number of new attributes that take care of the last
> few remaining areas that weren't configurable.  See the new attribute
> search_results_contenttype in this coming Sunday's 3.1.6 snapshot, as
> well as the existing add_anchors_to_excerpt attribute.

Ive got some CVS of htdig. So I'll update that.

> > What I would like to know: the summary strings you generate, are they
> > well-formed XML (in particular is there for each opening tag also a
> > closing tag? )
>
> You probably want to turn off anchors in excerpts (although they are
> well-formed HTML/XHTML), and all the rest is under control of the
> template files.  I think the distributed templates are all well-formed
> with closing tags.  Exceptions to this right now are the <br>, <hr>,
> <img>, <input> and <option> tags in the header, footer, nomatch, syntax
> and wrapper HTML files, but you're going to change them anyway.  Oh,
> there's also a non-self-terminating <br> tag in long.html and short.html
> that you may want to change to <br/> or remove.

Yes, I rolled myself some MRML.html, which generates query-result-element 
MRML tags.

The main problem I see is that I think a purely command line version of htdig 
would be cool. Purely commandline in the sense that you can write some shell 
that sais:

/path/to/htsearch --no-cgi --query-string="eggs and bacon" --sorted-by=score 
>> results.xml

(or some less fancy options style) and there is no interaction. Would you 
like me to start working on that some time? 

Currently, in fact, I have my ht_cli done (yes, a very ugly hack, and surely 
not commitable) and I am looking a bit more in the GIFT/popen side of things, 
which has me working on modifying Perl GIFT client for testing GIFT/HtDig.

To summarise: 
1) I will surely look deeper into templates (...and I will surely contribute 
the templates when they are done). 
2) I am waiting for your comments on adding some code that makes it possible 
to call htsearch by a batch, and I would also like to get rid of the http 
headers in the output(or did I overlook something??)
3) By simple modification of a couple of templates I seem to get the XML I 
want out of htsearch, at least for the case where there is no error (I tried 
that in the ten minutes between writing 2) and 3).

Cheers, and thanks,
Wolfgang

-- 
Dr. Wolfgang M&uuml;ller, assistant == teaching assistant
Personal page: http://cui.unige.ch/~vision/members/WolfgangMueller.html 
Maintainer, GNU Image Finding Tool (http://www.gnu.org/software/gift)